Adaptive AUTOSAR is considered to be by many automotive experts the new way of working when designing automotive software. This new standard can better cover the needs of high-performance computers and can tackle the requirements of our modern era.

In the early days of introducing ECUs (Electronic Control Units) in cars, each time the hardware was changed, due to reasons like the end of life of a component or a cost reduction due to a new, cheaper alternative, the entire software had to be rewritten from scratch. This has put high pressure on the cost and time to market, not only from the perspective of writing the code but also when considering the time to validate the new alternative.

A consortium of OEMs and Tier 1 suppliers came up with what is called Automotive Open Software Architecture abbreviated as AutOSAr. This new approach provides a base with very well-defined interfaces that allow different components to be ported from one chip to another and drastically lower the cost and time to market for the new ECUs.

As ECUs became more and more powerful, they started using Unix-like operating systems (mostly Linux and later Android) a new standard was needed to accommodate this trend. 

Understanding How This New Standard for high-performance ECUs works

Today, the advancements in technology have enabled specialists to design a new embedded standard for vehicle design, called Adaptive AUTOSAR. This improved standard enables more powerful and flexible in-vehicle E/E architectures and comes with several improvements over the standard AUTOSAR model.

In this post, we want to dive deeper into the specs of AUTOSAR adaptive and discover the mechanisms that drive the flexibility of the adaptive platform itself. Ultimately, we’ll show you what a framework of the Adaptive model would look like so that OEMs and suppliers could build improved, more flexible applications.

Let’s start with the three main automotive trends that are powering the growth of Adaptive AUTOSAR.

Automotive Trends Connected with Adaptive AUTOSAR

1. Mobility Concepts

Today, we’re experiencing a rise in mobility with changing customer behavior. We can notice an increase in car sharing, ride sharing, and upcoming mobility services, as well as a rise in vehicle subscription services, where you can order software right in your vehicle.

2. Electrification

In order to reduce emissions of today’s vehicles, we’re noticing an increase in governmental regulations. More and more governments are poised to drive the growth of electric vehicles and have a common goal to increase vehicle efficiency while minimizing pollution. This shift towards driving is being actively promoted inside the European Union, whose aim is to completely ban all diesel cars by 2050 and increase sales of electric cars by up to 30% until 2030.

We definitely need more intelligent vehicle-to-grid optimization for situations where many vehicles are returning to their residential neighborhoods and starting charging batteries simultaneously. This can be accomplished using smart IoT ecosystems.

3. Automated Driving

Automated driving is different than autonomous driving. This concept involves using different sensor types, such as sonar, camera, or radar, while taking care of high data throughput. The on and off data has to be combined and evaluated by complex algorithms to get an optimal image of the surroundings and ensure the driver stays safe behind the wheel.

Mobility as a Service

It seems like all these user experiences are preparing disruptive changes, triggering a new potential scenario: mobility as a service (MaaS). This concept involves the integration of various forms of transportation means into a single mobility service that is accessible on demand. The aim of MaaS is to enable people and businesses alike to be able to utilize various transportation vehicles, such as cars, taxis, scooters, metros, busses, or bikes, in order to travel in a city or across a larger area.

One of the biggest impacts of this service is the shift away from car ownership. This means less pollution overall and less crowded cities, which is exactly what governments need at this moment.

Enablers for Adaptive AUTOSAR

Adaptive AUTOSAR is driven by MaaS, automated driving, electrification, and mobility concepts. However, all its potency revolves around enablers. One of those enablers is connectivity. The vehicle is always on, being connected at all times to the infrastructure and to the other cars around. Another enabler is the offboard ecosystems, which are offering cloud services for data collection and for downloading high-resolution maps and road information.

One of the most interesting enablers for Adaptive AUTOSAR is the high-performance computing resources (HPCs) on board. HPCs consist of integrated multiple computing units that work together to improve the vehicle’s functionality. 

Going from AUTOSAR Classic to AUTOSAR Adaptive

Adaptive AUTOSAR focuses not only on adding more software to the car but also on creating a complete infrastructure from development to deployment. In 2022, the challenge for automakers is to bridge the gap between a deeply embedded system (vehicle) and the necessary back-end infrastructures.

While AUTOSAR Classic only focuses on embedded systems, its Adaptive counterpart focuses on embedded integration & debugging, security, safety, the automotive supply chain, communication protocols, and diagnostics. AUTOSAR Adaptive involves using onboard supercomputers that are offering their computing power for infotainment, head unit, and automated driving.

The supercomputers inside modern cars can be mastered by the OEMs so that the OEMs can install new software packages on the ECUs wirelessly. Of course, this doesn’t mean that Classic AUTOSAR will just go away. Instead, it will still continue to be used inside vehicles and will be responsible for controlling intelligent sensors and actuators, with the purpose of providing basic functions within the car. 

If we compared the two systems on a higher level, we would notice that AUTOSAR Classic has a monolithic structure with static connections that are optimized for efficiency, while its Adaptive counterpart has a dynamic structure, with a modular architecture that is optimized for flexibility. AUTOSAR Adaptive is the right choice even for smaller ECUs, where good software design is more important than having a small footprint.

In terms of vehicle architecture, Adaptive AUTOSAR can complement the Classic platform, and can also function alongside other commercial off-the-shelf services, which are proprietary infotainment solutions.

The Architecture of AUTOSAR Adaptive

AUTOSAR Adaptive comes with an improved architecture over its Classic counterpart, using C++ as a programming language and utilizing AUTOSAR Runtime as a base for apps. It uses incremental deployment and enables the developers to change the configuration during runtime.

While Classic AUTOSAR is static in nature, Adaptive AUTOSAR has a “planned dynamics” appearance, both in communication, resources, and application deployment.

The Bottom Line

Adaptive AUTOSAR is a modern architecture that is not designed to replace the Classic AUTOSAR but to build on top of it. This concept comes alive in 2022 as new resources and requirements are coming to the automotive industry, including vehicle connectivity, autonomous driving, and updates. If you want to learn more about Tremend’s take on Adaptive AUTOSAR and our experience in similar projects, you can visit our dedicated page.